Obesity and Ovarian Cancer Survival: A Systematic …cancerpreventionresearch.aacrjournals.org/content/canprevres/5/7/...Obesity and Ovarian Cancer Survival: A Systematic Review and

Review

Obesity and Ovarian Cancer Survival: A Systematic Reviewand Meta-analysis

Melinda M. Protani1,2, Christina M. Nagle1, and Penelope M. Webb1

AbstractStudies that have examined the association between obesity and ovarian cancer survival have provided

conflicting results. We reviewed and quantitatively summarized existing evidence, exploring potentially

important sources of variability, such as the timing of body mass index (BMI) assessment and different

cutpoints used to categorize BMI. A systematic search of MEDLINE and EMBASE was conducted to identify

original data evaluating the association between obesity and survival in women with ovarian cancer.

Adjusted hazard ratios (HR) from studies were pooled using a random-effects model. The meta-analysis of

14 studies showed slightly poorer survival among obese than in non-obese women [pooled HR, 1.17; 95%

confidence interval (CI), 1.03–1.34]. This estimate did not vary appreciably whenBMIwasmeasured before

diagnosis (1.13; 0.95–1.35), at the time of diagnosis (1.13; 0.81–1.57) or at the commencement of

chemotherapy (1.12; 0.96–1.31). We found a slightly stronger association in studies that only included

womenwith a BMI� 30 in their "obese" group (1.20) than in studies that also included overweight women

(BMI � 25; 1.14). Women with ovarian cancer who are obese appear to have slightly worse survival than

non-obese women. However, there is a large amount of inter-study variation, which means that no solid

conclusions can be drawn. Cancer Prev Res; 5(7); 901–10. �2012 AACR.

IntroductionOvarian cancer is a highly fatal disease, with only about

40% of women with ovarian cancer still alive more than 5years postdiagnosis (1). This poor survival is largely attrib-utable to the fact that approximately 75% of all cases ofovarian cancer in developed countries are diagnosed withmetastatic spread beyond the pelvis (1, 2). While stage ofdisease at diagnosis remains the most important predictorof survival time, other knownprognostic factors include ageat diagnosis, tumor grade, and the amount of residualdisease following surgery (3, 4). However, at the time ofdiagnosis, none of these factors are amenable to interven-tion to improve survival.Potentially modifiable factors such as obesity, com-

monly measured by body mass index (BMI), have beenfound to be associated with poorer survival in a numberof cancers including breast (5), prostate (6), and colo-rectal cancer (7). Few studies have examined the associ-ation between obesity and ovarian cancer survival andthose that have provided conflicting results. Furthermore,

it is unclear whether sources of heterogeneity betweenstudies, such as the timing of BMI assessment or the cutoffpoints used to classify BMI, may be contributing to thesediscrepancies.

A recent meta-analysis (8) of studies published up toDecember 2010 found that women with ovarian cancerwho were obese during early adulthood (3 studies) orbefore diagnosis had worse survival (5 studies); however,no association with obesity measured around diagnosis (5studies). Currently, it is unclear whether BMI in earlyadulthood or before diagnosis, the focus of the previousmeta-analysis, is the relevant biologic window. For exam-ple, the practice of chemotherapy dose capping in obesepatients (to prevent toxicity) may have negative implica-tions on survival outcomes (9), so body size at the com-mencement of chemotherapy may be more relevant. Sincethis previous meta-analysis, there have been a number ofadditional epidemiologic studies published on the associ-ationbetweenBMI andovarian cancer survival, andwehavealso identified additional studies that were not included inthe previous meta-analysis (10–15).

Given the growing number of studies in the literature andincreasing interest in the role of lifestyle factors in cancersurvival, our aim was to systematically re-evaluate theliterature examining the association between obesity andsurvival in women with ovarian cancer and to conduct anupdated, more comprehensive meta-analysis to quantifythe magnitude of risk. A second specific objective was toexplore potentially important sources of variability, such asthe timing of BMI assessment and the different cutoff pointsused to categorize BMI.

Authors' Affiliations: 1Gynaecological Cancers Group, Queensland Insti-tute ofMedical Research; and 2School of Population Health, TheUniversityof Queensland, Herston, Queensland, Australia

Corresponding Author: Melinda M. Protani, Gynaecological CancersGroup, Queensland Institute of Medical Research, Locked Bag 2000 RoyalBrisbane Hospital, Herston, QLD 4029, Australia. Phone: 61-733620226;Fax: 61-738453502; E-mail: [email protected]

doi: 10.1158/1940-6207.CAPR-12-0048

�2012 American Association for Cancer Research.

CancerPreventionResearch

www.aacrjournals.org 901

Research. on July 3, 2018. © 2012 American Association for Cancercancerpreventionresearch.aacrjournals.org Downloaded from

Published OnlineFirst May 18, 2012; DOI: 10.1158/1940-6207.CAPR-12-0048

http://cancerpreventionresearch.aacrjournals.org/

Materials and MethodsSearch strategy

This systematic review and meta-analysis was conductedaccording to the Meta-analysis of Observational Studies inEpidemiology (MOOSE) guidelines (16). A systematicsearch of MEDLINE and EMBASE, from inception to Sep-tember 2011, was conducted to identify studies examiningthe association betweenobesity and survival inwomenwithovarian cancer. The search included terms for ovariancancer (ovarian neoplasms OR ovarian cancer OR ovariantumorORovarian tumourORovarian carcinoma)ANDobesity(body size OR body weight OR overweight OR obesity OR bodymass index) AND survival (survival analysis OR survival rateOR proportional hazards model OR survival OR prognosis). Thereference lists of all eligible articles and reviews were alsoscanned to identify additional studies for inclusion.

Study selection and data extractionStudies were eligible for inclusion in the systematic

review if they contained original data examining the asso-ciation between obesity (assessed by any measure) andsurvival (ovarian cancer–specific survival or overall surviv-al) in a cohort of women newly diagnosed with ovariancancer. To be eligible for inclusion in the meta-analysis,studies had to additionally provide hazard ratio (HR)estimates. For all eligible studies, information was extractedon study design, country, years of diagnosis, years of follow-up, age, stage, definitions and categories of BMI, the timingof when BMI was measured, median survival, effect esti-mates, and variables adjusted for in analyses. Where morethan one HR was reported, the most fully adjusted HR wasextracted for the meta-analysis.

Statistical analysisHR estimates were pooled using random-effects meta-

analysis (17), and the heterogeneity across studies was

assessedusing the I2 statistic (18). Studies examiningoverallsurvival were pooled with studies examining ovarian can-cer–specific survival as previous research has shown thatthere are very few competing causes of death in this pop-ulation of women due to the highly fatal nature of ovariancancer (1). Where multiple measurements of obesity weretaken throughout the course of the cancer (e.g., from pre-diagnosis through to the commencement of chemothera-py), the estimate closest to body weight before diagnosiswas used for primary analyses as most studies examinedprediagnosis body weight. The majority of studies (n ¼ 9)reported estimates for categories of BMI, similar to that ofthe World Health Organization guidelines (19). For the 2studies that reported the effect of BMI as a continuousvariable (14, 20), we used the reported effect sizes and95% confidence intervals (CI) per 1-unit increase in BMIto estimate the HR and corresponding 95% CIs for a 5-unitchange inBMI for comparabilitywith the estimates reportedin other studies.

Prespecified sensitivity analyses were conducted to assesswhether there was a differential effect on survival accordingto when obesity was measured (before diagnosis, at diag-nosis, or at chemotherapy) as well as the definition ofobesity used for analysis (BMI � 30, BMI � 25 or per 5-unit increase in BMI). Publication bias was assessed byexamining funnel plot asymmetry (21, 22). All analyseswere conducted using Stata 11.0 (23).

ResultsSystematic review

The primary search identified 57 eligible titles. Afterreview of the abstracts, we identified 20 studies that wereeligible for inclusion in the systematic review (Fig. 1and Table 1). The 20 studies included women diagnosedwith ovarian cancer between 1977 and 2007 with cohortsfrom the United States (n¼ 11), Sweden (n¼ 2), Germany

57 eligible titles

20 eligible abstracts included in the systematic review

15 unsuitable – exclusion A

10 unsuitable – exclusion B

2 unsuitable – exclusion C

8 unsuitable – exclusion D

2 unsuitable – exclusion E

Abstracts

screened for

eligibility

462 results from

MEDLINE, CINAHL and reference lists

20 articles

assessed for

eligibility for

meta-analysis

Total articles included in meta-analysis = 14

Figure 1. Study selection: exclusioncriteria for the systematic review.Studies which did not evaluate aprognostic outcome (recurrence,disease-free survival, progression-free survival, all-cause mortality, orovarian cancer–specific survival/mortality) in ovarian cancerpatients (A); did not report originaldata (B); examined possiblemolecular pathways for obesity-related cancer survival (C); did notassess obesity status or did notanalyze the effect of obesity onovarian cancer prognosis (D);and contained overlappingpopulations (E).

Protani et al.

Cancer Prev Res; 5(7) July 2012 Cancer Prevention Research902




Tab

le1.

Cha

racteristic

sof

stud

iesex

aminingtheas

sociationbetwee

nob

esity

andlong

-term

outcom

esin

patientswith

ovarianca

ncer

Source

(coun

try)

NYea

rsof

diagno

sis

Follo

w-up,

yAge,

yStage

Exp

osu

re(BMIc

ateg

ory)

Med

ian

survival

time,

mo

HR

(95%

CI)

Adjustmen

tva

riab

les

Obse

rvationa

lco

horts

Dolec

ekan

dco

lleag

ues

(USA;ref.1

3)

341

1994

–19

98Max

imum

,11

Ran

ge,1

8–74

All

18.5–24

.9b

�30

1.0

1.20

(0.72–

1.98

)Age

,stage

,grade

,rac

e,residua

lles

ions

,sm

oking,

OC

use,

parity

Fotopou

lou

andco

lleag

ues

(German

y;ref.10

)

306

2000

–20

10Med

ian,

0.97

Ran

ge,0

.01–

5.2

Ran

ge,1

8–92

All

<25b

�25

1.0

0.73

(0.39–

1.37

)Age

,stage

,grade.

lymphno

destatus

,residua

ltum

our,

ascites,

IMO

leve

linvo

lvem

ent,

nons

erou

shistolog

y,distant

metas

tase

sKjaerbye

-Th

yges

enan

dco

lleag

ues

(Den

mark;

ref.27

)

295

1994

–19

99Med

ian,

7.3

Ran

ge,5

.4–9.5

Ran

ge,3

5–79

III18

.5–24

.9a

�25

33.6

25.2

1.0

1.83

(1.38–

2.42

)Age

,rad

icality

ofsu

rgery,

histolog

y,platin

um-bas

edch

emothe

rapy

,sm

oking

Lamkinan

dco

lleag

ues

(USA;ref.1

4)

7420

01–20

05Mea

n,2.01

Ran

ge,0

.02–

6.08

Med

ian,

62Ran

ge,3

3–87

All

Per

1-un

itb

increa

sein

BMI

1.01

(0.97–

1.04

)Nil

Matthew

san

dco

lleag

ues

(USA;ref.3

4)

304

1996

–20

05Max

imum

,10

BMI<

30(m

ean,

62.2)

BMI�

30(m

ean,

58.3)II–IV

18.5–24

.9e

�35

40 48 P¼

0.37

——

Moy

sich

and

colleag

ues

(USA;ref.2

8)

359

1982

–19

98Minim

um,9

Mea

nalive,

47.5

Mea

ndea

d,5

8.3

All

<25a

�30

57 591.0

0.99

(0.71–

1.38

)Age

,stage

Mun

sted

tan

dco

lleag

ues

(German

y;ref.31

)

824

1986

–20

05Med

ian,

5.13

Med

ian,

60.5

All

20–25

b

30–40

20.28

23.04

Ptrend¼

0.05

3

——

Nag

lean

dco

lleag

ues

(Aus

tralia;ref.2

9)

609

1990

–19

93Mea

n,7.3

Ran

ge,5

–8.3

Ran

ge,1

8–79

All

<22.2a

�25.8

1.0

0.96

(0.74–

1.23

)Age

,stage

,grade,

total

energy

intake

,residua

l,as

cites,

smok

ing,

parity

,OC

use

(Con

tinue

don

thefollo

wingpag

e)

Obesity and Ovarian Cancer Survival: Meta-analysis

www.aacrjournals.org Cancer Prev Res; 5(7) July 2012 903




Tab

le1.

Cha

racteristic

sof

stud

iesex

aminingtheas

sociationbetwee

nob

esity

andlong

-term

outcom

esinpatientswith

ovarianca

ncer

(Con

t'd)

Source

(coun

try)

NYea

rsof

diagno

sis

Follo

w-up,

yAge,

yStage

Exp

osu

re(BMIca

tegory)

Med

ian

survival

time,

mo

HR

(95%

CI)

Adjustmen

tva

riab

les

Pav

elka

and

colleag

ues

(USA;ref.2

0)

149

1996

–20

03Not

stated

Ran

ge,1

8–79

III–IV

Per

1-un

itc

increa

sein

BMI

18.5–24

.9�3

080 62 P¼

0.28

1.05

(1.005

–1.09

7)Nil

Sch

lumbrech

tan

dco

lleag

ues

(USA;ref.1

5)

127

2002

–20

07Mea

n,3.1

Ran

ge,0

.3–7.2

Not

stated

All

Not

stated

c1.0

0.95

(0.68–

2.43

)Not

stated

Sch

lumbrech

tand

colleag

ues

(USA;ref.1

1)

194

1977

–20

09Med

ian,

5.1

Ran

ge,0

.1–31

.9Mea

n,44

.9Ran

ge,1

4–79

All

<25c

�30–

<35

�35

1.0

1.02

(0.43–

2.38

)2.53

(1.19–

5.38

)

Nil

Sch

ildkrau

tan

dco

lleag

ues

(USA;ref.3

0)

257

1980

–19

82Med

ian,

8.3

Ran

ge,0

.1–14

.2Mea

n,43

.7Ran

ge,2

0–54

All

<27.9a

�27.9

52 64 P¼

0.51

1.0

1.1(0.7–1.7)

Age

,stage

,p53

status

Skirnisdo

ttiran

dco

lleag

ues

(Swed

en;ref.3

2)

446

1994

–20

03Mea

n,3.9

Ran

ge,0

–12

.3Mea

n,62

.5All

�25c

>25

1.0

0.94

(0.74–

1.21

)Age

,stage

,histology

Suh

and

colleag

ues

(Korea

;ref.3

3)

486

2000

–20

10Med

ian,

2.83

Ran

ge,0

–13

.2BMI�

23(m

ean,

53.2)

BMI<

23(m

ean,

48.6)

All

<23c

�23

P¼

0.67

——

Yan

gan

dco

lleag

ues

(Swed

en;ref.2

6)

635

1993

–19

95Not

stated

Ran

ge,5

0–74

All

18.5–24

.9a

�30

1.0

1.22

(0.86–

1.71

)Age

,stage

,grade

Zha

ngan

dco

lleag

ues

(China

;ref.3

5)

207

1999

–20

00Minim

um,3

Mea

nalive,

46.7

Mea

ndea

d,51

.6All

<20

�25a

�20b

�25

1.0

2.33

(1.12–

4.87

)1.0

0.76

(0.38–

1.52

)

Age

,stage

,grade

,as

cites,

residu

allesion

s,ch

emothe

rapy,

total

energy

intake

,men

opau

sals

tatus

Zho

uan

dco

lleag

ues

(USA;ref.1

2)

388

1998

–20

03Max

imum

,5Mea

n,58

.6All

<25a

�25

<25d

�25

1.0

1.30

(0.92–

1.83

)1.0

1.05

(0.75–

1.48

)

Age

,stage

,histology

,ed

ucation,

OC

use,

men

opau

sals

tatus,

HRTus

e,parity

,age

atfirstbirth,

family

historyof

ovarian

canc

er,tim

efrom

diagn

osis

tostud

y

(Con

tinue

don

thefollo

wingpag

e)

Protani et al.





Tab

le1.

Cha

racteristic

sof

stud

iesex

aminingtheas

sociationbetwee

nob

esity

andlong

-term

outcom

esinpatientswith

ovarianca

ncer

(Con

t'd)

Source

(coun

try)

NYea

rsof

diagno

sis

Follo

w-up,

yAge,

yStage

Exp

osu

re(BMIca

tegory)

Med

ian

survival

time,

mo

HR

(95%

CI)

Adjustmen

tva

riab

les

Treatmen

tco

horts

Barrettan

dco

lleag

ues

(multip

leco

untries;

ref.24

)

1,06

719

98–20

00Not

stated

Med

ian,

59Ran

ge,1

9–85

IC–IV

18.5–24

.9c

�30

Not

attained

34.3

P¼

0.10

——

Hes

san

dco

lleag

ues,

(USA;ref.2

5)

790

1995

–19

98Med

ian,

4Ran

ge,2

1–90

III<2

5c

�30

54.3

48.4

——

p¼

0.62

Wrig

htan

dco

lleag

ues

(USA;ref.9

)

387

Not

stated

Med

ian,

4.4

Med

ian,

56.8

Ran

ge,2

1–85

Not

stated

"Acros

sBMI

strata"c

P¼

0.41

——

Abbreviations

:HRT,

horm

onereplace

men

ttherap

y;OC,o

ralc

ontrac

eptiv

e.aBMIm

easu

redbeforediagn

osis.

bBMIm

easu

redat/aroun

dtim

eof

diag

nosis.

cBMIm

easu

redat

theco

mmen

cemen

tof

chem

othe

rapy.

dBMIm

easu

red9mon

thspo

st-che

mothe

rapy.

eTimeof

BMIm

easu

remen

tno

tstated

.






(n¼ 2), Denmark (n¼ 1), Australia (n¼ 1), China (n¼ 1),Korea (n ¼ 1), and 1 cohort involving multiple countries.Sample size ranged from 74 to 1,067, with amedian of 350.The majority of the studies were observational cohorts;however, 3 were cohorts of women with ovarian cancerparticipating in randomized trials (9, 24, 25).

All studies used BMI as ameasure of obesity; however, thetime point when BMI was measured, as well as the cutoffpoints used to categorize BMI for analysis varied betweenstudies. Five studies used data on height and weightobtained 1 year before diagnosis (26) or from reports ofwomen’s usual adult weight (27–30), 4 studies measuredBMI at the time of diagnosis (10, 13, 14, 31), 8 at thecommencement of chemotherapy (9, 11, 15, 20, 24, 25, 32,33), 1 study did not statewhenBMIwasmeasured (34), and2 studies assessed BMI at multiple time points (12, 35)including 5 years before diagnosis (12, 35). The cutoffpoints used to categorize the obese group were in accordwith the World Health Organization’s International Clas-sification of Obesity in approximately half of the studies(BMI � 30 kg/m2 being obese; n ¼ 9; ref. 19). However, 8studies used a combined overweight/obese group (BMI �25 kg/m2), 2 studies analyzed their data per 1-unit increasein BMI, 2 studies analyzed data as semicontinuous variablesacross BMI strata, and 1 study did not state how BMI wascategorized for analysis. Five studies used the World HealthOrganization’s classification of normal BMI (18.5–24.9 kg/m2) as the reference group (13, 24, 26, 27, 34) whereasothers used variations including all womenwith a BMI < 20or BMI < 25. Median follow-up time varied considerablybetween studies ranging from less than1 year to greater than10 years. Thirteen studies used all-cause mortality as the

endpoint, whereas 7 studies used ovarian cancer–specificdeaths as the endpoint. Nine of the studies adjusted for thekey prognostic factors of stage at diagnosis and age, otherprognostic factors were adjusted for less consistently.

Three observational cohorts (31, 33, 34) and the 3treatment cohorts (9, 24, 25) did not report HRs and sowere not included in our initial meta-analysis. All of thesestudies reported that survival time did not differ significant-ly between BMI strata, with the exception of the study byMunstedt and colleagues, which found a trend towardimproved survival in women who were obese (31). Esti-mates for 2 of these studies (25, 31) were, however, includ-ed in the previous meta-analysis (8), thus we conducted asensitivity analysis including this additional information.

Meta-analysisOur meta-analysis of the 14 studies showed slightly

poorer survival among the obese group compared withnon-obese women with ovarian cancer [pooled HR (pHR),1.17; 95% CI, 1.03–1.34; Fig. 2]. This estimate did not varyappreciably when we restricted it to studies where BMI wasmeasured before diagnosis (pHR, 1.13; 0.95–1.35), at thetime of diagnosis (pHR, 1.13; 0.81–1.57), or at the time ofchemotherapy (pHR, 1.13; 0.92–1.39; Fig. 3). There was alarge amount of inter-study heterogeneity among the BMIcutoff points used to define both the "obese" group and the"reference" group for analysis. The survival differentialvaried only slightly depending on whether the "obese"group included only women with a BMI � 30 (pHR,1.20; 95% CI, 0.94–1.53), obese and overweight women(BMI � 25; pHR, 1.14; 95% CI, 0.92–1.41), or whetherresults were analyzed per 5-unit increase in BMI (pHR, 1.15;

Overall (I 2 = 51.1%, P = 0.012)

Schlumbrecht 2009 (15)

Pavelka 2006 (20)

Zhang 2005 (35)

Fotopoulou 2011 (10)

Nagle 2003 (29)

Schlumbrecht 2011a (11)

Dolecek 2010 (13)

Lamkin 2009 (14)

Zhou 2011 (12)

Kjaerbye-Thygesen 2006 (27)

Skirnisdottir 2010 (32)

Schildkraut 2000 (30)

Schlumbrecht 2011b (11)

Yang 2008 (26)

Moysich 2007 (28)

Study ID

1.17 (1.03–1.34)

0.95 (0.68–2.43)

1.28 (1.03–1.59)

2.33 (1.12–4.87)

0.73 (0.39–1.37)

0.96 (0.74–1.23)

1.02 (0.43–2.38)

1.20 (0.72–1.98)

1.05 (0.86–1.22)

1.30 (0.92–1.83)

HR (95% CI)

1.83 (1.38–2.42)

0.94 (0.74–1.21)

1.10 (0.70–1.70)

2.53 (1.19–5.38)

1.22 (0.86–1.71)

0.99 (0.71–1.38)

1.5 1 5 10

HR (95% CI; log scale)

Figure 2. Meta-analysis and pHR ofthe effect of obesity on survival inpatients with ovarian cancer. Note:Schlumbrecht 2011a: BMI ¼ 30–35; Schlumbrecht 2011b: BMI �35.

Protani et al.





95% CI, 0.95–1.39; Fig. 4). Because studies used differentmethods to account for confounding, we conducted a posthoc sensitivity analysis excluding all studies that did notadjust for at least age and stage (n¼ 5) and obtained a pHRof 1.17 (95% CI, 0.97–1.40). Inclusion of the estimates forthe2 additional studies (as reportedbyYangand colleagues;ref. 8) reduced the estimate slightly to 1.13 (95% CI, 1.01–1.28).

Publication biasThe funnel plot of the effect estimates of obesity and

ovarian cancer survival was close to symmetrical, and therewas no evidence of publication bias using the Egger weight-ed regression method (Pbias ¼ 0.44) or the Begg rankcorrelation method (Pbias ¼ 0.32).

DiscussionIn this meta-analysis, we have found consistent evidence

that survival among obese women with ovarian cancer isslightly worse than survival among non-obese women. Onthe basis of our analysis of the published literature, weestimate that the risk of survival among obese women withovarian cancer is 15% to 20% worse than women with aBMI in the "healthy" range. Our results were consistentregardless of whether BMI was measured before diagnosis,

at diagnosis, or at/around the commencement of chemo-therapy. Compared with the previous meta-analysis, oursummary estimate is larger for obesity measured at oraround the time of diagnosis (pHR, 1.13 vs. 0.94; ref. 8).This is, in part, due to the different criteria used to defineobesity at or around the time of diagnosis and hence theinclusion of different studies in the 2 pooled calculations.The other major difference between our meta-analysis andthe previous meta-analysis was the HR from one of thestudies. The studyby Pavelka and colleagues reported anHRof 1.05 per 1-unit increase in BMI (20), so for consistencywith the other studies in our meta-analysis, we convertedthis estimate to give an expected HR of 1.28, for a 5-unitincrease in BMI. These estimates contrast markedly, how-ever, with the HR of 0.53 that Yang and colleagues includedin their meta-analysis (8).

Our meta-analysis also adds to the previous analysis inthat it explored several potentially important sources ofinter-study variability. One such source of variation is theBMI cutoff points used to classify the obese and referencegroups for analysis. Inclusion of underweight women (whoare likely to have worse outcomes) in the reference groupand/or overweight women in the obese group may under-estimate the true association between obesity and ovariancancer survival. Our sensitivity analysis, which stratifiedstudies by how they defined obesity, suggested that there

Figure 3. Sensitivity analyses ofpHRs of the effect of obesity onsurvival in patients with ovariancancer, stratified by the timing ofwhen obesity was measured. Note:Schlumbrecht 2011a: BMI ¼ 30–35;Schlumbrecht 2011b: BMI � 35.

.

.

.

1. Before diagnosis


Nagle 2003 (29)

Zhang 2005 (35)

Moysich 2007 (28)

Yang 2008 (26)

Zhou 2011 (12)

Subtotal (I2 = 25.6%, P = 0.242)

2. At diagnosis

Zhang 2005 (35)


Lamkin 2009 (14)

Dolecek 2010 (13)


Subtotal (I2 = 73.1%, P = 0.005)

3. At/around chemotherapy

Pavelka 2006 (20)

Schlumbrecht 2009 (15)



Zhou 2011 (12)


Subtotal (I2 = 39.9%, P = 0.139)

Study ID

1.10 (0.70–1.70)

0.96 (0.74–1.23)

2.33 (1.12–4.87)

0.99 (0.71–1.38)

1.22 (0.86–1.71)

1.30 (0.92–1.83)

1.13 (0.95–1.35)

0.76 (0.38–1.52)

1.83 (1.38–2.42)

1.05 (0.86–1.22)

1.20 (0.72–1.98)

0.73 (0.39–1.37)

1.13 (0.81–1.57)

1.28 (1.03–1.59)

0.95 (0.68–2.43)

0.94 (0.74–1.21)

1.02 (0.43–2.38)

1.05 (0.75–1.48)

2.53 (1.19–5.38)

1.13 (0.92–1.39)

HR (95% CI)

1.5 1 5 10







was a slightly stronger effect in studies that only includedwomenwith a BMI� 30 in their "obese" group (pHR, 1.20)than in studies that also included overweight women (BMI� 25; pHR, 1.14). We also identified a large amount ofvariability about the time point when BMI was measured.Changes in weight and body composition commonly occurthroughout the course of ovarian cancer. Both weight loss,generally due to cachexia, and weight gain, typically due toascites, can be presenting symptoms for ovarian cancer,particularly in women with advanced disease (36). Weightchange can also occur during treatment and is likely to beassociated with outcome (weight gain being an indicator ofimproved survival and weight loss an indicator for poorsurvival; ref. 25). The timing of BMI measurement is there-fore particularly important as it determines the specificresearch questions being asked.

First, women who are obese before, or at diagnosis, mayhavemorebiologically aggressive tumors as excess adiposityis associated with the upregulation of a number of cellularproliferation pathways which may lead to increased tumorgrowth and metastasis (37). For example, leptin, an adipo-cytokine produced by white adipose tissue, is known to actas a growth factor in a number of cancer cell lines includingbreast, endometrial, and prostate cancers (38, 39) and isalso involved in promoting angiogenesis (40).

Second, chemotherapy dosage is calculated on the basisof body surface area. Because of concerns of relative

overdosing in obese patients with a large body surfacearea, it is well documented that empiric dose capping ofchemotherapeutic drugs (usually at a body surface area ofeither 1.8 or 2 m2) occurs in some centers (41). Further-more some, but not all, observational studies have shownthat dose intensity (42) and the cumulative dose (20) ofchemotherapy may be lower in obese women (comparedwith normal weight). Evidence also suggests that obesewomen with ovarian cancer who have their doses cappedat 2.0 m2 experience similar or lower rates of chemother-apy-induced toxicities compared with those who weredosed according to their actual body weight, a furtherindication that obese women may be receiving subopti-mal treatment, and therefore be at an increased risk ofdisease progression and reduced survival (9,43). Obesityis also associated with other comorbidities such as dia-betes and cardiovascular disease, which may also lead towomen being treated with reduced doses of chemother-apy (44), as well as being independently associated withoverall survival. The potential role of reverse causation(where deteriorating health status may influence bodysize) also needs to be considered.

Interestingly, in our sensitivity analysis, the associationbetween obesity and survival did not appear to vary appre-ciably by whether a woman’s obesity status was measuredbefore diagnosis, at diagnosis, or at the time of chemother-apy. However, the paucity of published data in relation to

.

.

.

1. Obese only

Moysich 2007 (28)

Yang 2008 (26)

Dolecek 2010 (13)



Subtotal (I2 = 22.3%, P = 0.272)

2. Obese + overweight


Nagle 2003 (29)

Zhang 2005 (35)

Zhang 2005 (35)



Zhou 2011 (12)


Zhou 2011 (12)

Subtotal (I2 = 64.8%, P = 0.004)

3. Per 5-unit increase

Pavelka 2006 (20)

Lamkin 2009 (14)

Subtotal (I2 = 48.4%, P = 0.164)

Study ID

0.99 (0.71–1.38)

1.22 (0.86–1.71)

1.20 (0.72–1.98)

1.02 (0.43–2.38)

2.53 (1.19–5.38)

1.20 (0.94–1.53)

1.10 (0.70–1.70)

0.96 (0.74–1.23)

0.76 (0.38–1.52)

2.33 (1.12–4.87)

1.83 (1.38–2.42)

0.94 (0.74–1.21)

1.30 (0.92–1.83)

0.73 (0.39–1.37)

1.05 (0.75–1.48)

1.14 (0.92–1.41)

1.28 (1.03–1.59)

1.05 (0.86–1.22)

1.15 (0.95–1.39)

HR (95% CI)

1.5 1 5 10


Figure 4. Meta-analysis and pHRsof the effect of obesity on survival inpatients with ovarian cancerstratified by the cutoff points usedto define obesity in analyses:Obese-only (BMI � 30) versusobese and overweight (BMI � 25).Note: Schlumbrecht 2011a: BMI ¼30–35; Schlumbrecht 2011b: BMI� 35.

Protani et al.





differences in the timing of BMI measurement and associa-tionswith ovarian cancer survival limit conclusions that canbe drawn. Therefore, future studies should include carefulplanning of the timing of obesity measurement to elucidatethe causal mechanisms surrounding adverse survival inobese women with ovarian cancer.

Implications for further researchDifferences in dosing protocols for obese women may

explain someof the disparities seen in the results of differentstudies in thismeta-analysis; however, few studies providedinformation on dosing protocols. Future studies shouldideally specify dosing protocols, such as the percentage ofwomen receiving chemotherapy dose reductions, to help ininterpreting their results.To date, no studies have examined other measures of

obesity, such as waist–hip ratio (WHR), which has beenshown to be associated with reduced survival in womenwith breast cancer (45, 46). WHR considers the anatomicdistribution of adipose tissue, which is a more accurateindicator of metabolic stress associated with increased adi-posity, particularly when compared with BMI, which isunable to distinguish lean muscle mass from fat mass(47–49). In addition, as obesity appears to be differentiallyassociated with the incidence of ovarian cancer in pre- andpostmenopausal women and with different histologic sub-types of cancer (50, 51), future large-scale studies andpooled cohorts should aim to assess whether there is adifferential effect of obesity on survival according to thesefactors as well as other prognostic factors.Strengths of our review are the broad search strategy and

that references from all included studies and relevant nar-rative reviews were cross-checked for additional publica-tions. However, as with any meta-analysis, any biases andconfounding inherent in the original studies will also bepresent in our analyses (52). We have attempted to mini-mize the effect of confounding by using the most adjustedestimates provided by studies. Our sensitivity analysis,

which excluded studies that did not adjust (or restrict) forat least stage and age, suggested that the association betweenobesity and ovarian cancer survival was robust to potentialconfounding.

ConclusionThe results of our meta-analysis, based on more studies

thanprevious reviews, suggest that obesity is associatedwitha weak adverse effect on the survival of womenwith ovariancancer. However, the large amount of inter-study hetero-geneity means that no firm conclusions can be drawn.Further studies need to be conducted with a particular focuson selecting the timing of themeasurement of obesity basedon specific mechanistic hypotheses such as the role ofrelative underdosing of chemotherapy.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: M.M. Protani, C.M. Nagle, P.M. WebbDevelopment of methodology: M.M. Protani, C.M. NagleAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): M.M. ProtaniAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): M.M. Protani, P.M. WebbWriting, review, and/or revision of the manuscript: M.M. Protani, C.M.Nagle, P.M. WebbAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): M.M. ProtaniStudy supervision: C.M. Nagle, P.M. Webb

Grant SupportM.M. Protani is funded by an Australian Postgraduate Award Scholarship.

C.M. Nagle and P.M. Webb are funded by Fellowships from the NationalHealth and Medical Research Council (NHMRC) of Australia.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received February 2, 2012; revised April 3, 2012; accepted May 1, 2012;published OnlineFirst May 18, 2012.

References1. Australian Institute of Health and Welfare & National Breast and

Ovarian Cancer Centre. Ovarian cancer in Australia: an overview,2010. Canberra, Australia: AIHW; 2010.

2. Howlader N, Noone AM, Krapcho M, Neyman N, Aminou R, WaldronW, et al. SEER cancer statistics review, 1975–2008. Bethesda, MD:National Cancer Institute.

3. Hoskins WJ, McGuire WP, Brady MF, Homesley HD, Creasman WT,Berman M, et al. The effect of diameter of largest residual disease onsurvival after primary cytoreductive surgery in patientswith suboptimalresidual epithelial ovarian carcinoma. Am J Obstet Gynecol1994;170:974–9; discussion 979–80.

4. Omura GA, BradyMF, Homesley HD, Yordan E, Major FJ, BuchsbaumHJ, et al. Long-term follow-up and prognostic factor analysis inadvanced ovarian carcinoma: the Gynecologic OncologyGroup expe-rience. J Clin Oncol 1991;9:1138–50.

5. Protani M, Coory M, Martin JH. Effect of obesity on survival of womenwith breast cancer: systematic review and meta-analysis. BreastCancer Res Treat 2010;123:627–35.

6. Cao Y,Ma J. Bodymass index, prostate cancer-specificmortality, andbiochemical recurrence: a systematic review and meta-analysis. Can-cer Prev Res 2011;4:486–501.

7. Vrieling A, Kampman E. The role of body mass index, physical activity,and diet in colorectal cancer recurrence and survival: a review of theliterature. Am J Clin Nutr 2010;92:471–90.

8. Yang HS, Yoon C, Myung SK, Park SM. Effect of obesity on survival ofwomen with epithelial ovarian cancer: a systematic review and meta-analysis of observational studies. Int J Gynecol Cancer 2011;21:1525–32.

9. Wright JD, Tian C, Mutch DG, Herzog TJ, Nagao S, Fujiwara K, et al.Carboplatin dosing in obese women with ovarian cancer: a Gyneco-logic Oncology Group study. Gynecol Oncol 2008;109:353–8.

10. FotopoulouC, Richter R, Braicu EI, KuhbergM, Feldheiser A, SchefoldJC, et al. Impact of obesity on operativemorbidity and clinical outcomein primary epithelial ovarian cancer after optimal primary tumor debulk-ing. Ann Surg Oncol 2011;18:2629–37.

11. SchlumbrechtMP,SunCC,WongKN, BroaddusRR,GershensonDM,Bodurka DC. Clinicodemographic factors influencing outcomes inpatients with low-grade serous ovarian carcinoma. Cancer 2011;117:3741–9.

12. ZhouY, IrwinML,RischHA. Pre- andpost-diagnosis bodymass index,weight change, and ovarian cancer mortality. Gynecol Oncol 2011;120:209–13.






13. Dolecek TA, McCarthy BJ, Joslin CE, Peterson CE, Kim S, Freels SA,et al. Prediagnosis food patterns are associated with length of survivalfrom epithelial ovarian cancer. J Am Diet Assoc 2010;110:369–82.

14. Lamkin DM, Spitz DR, Shahzad MMK, Zimmerman B, Lenihan DJ,Degeest K, et al. Glucose as a prognostic factor in ovarian carcinoma.Cancer 2009;115:1021–7.

15. Schlumbrecht M, Urbauer D, Gershenson D, Broaddus R. Prognosticsignificance of obesity in high-grade serous carcinoma of the ovary. JClin Oncol 2009;27:e16528.

16. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D,et al. Meta-analysis of observational studies in epidemiology: a pro-posal for reporting. Meta-analysis of Observational Studies in Epide-miology (MOOSE) group. JAMA 2000;283:2008–12.

17. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control ClinTrials 1986;7:177–88.

18. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring incon-sistency in meta-analyses. BMJ 2003;327:557–60.

19. WHO. Obesity: preventing and managing the global epidemic. Reportof a WHO Consultation. Geneva, Switzerland: World Health Organi-zation; 2000.

20. Pavelka JC, Brown RS, Karlan BY, Cass I, Leuchter RS, Lagasse LO,et al. Effect of obesity on survival in epithelial ovarian cancer. Cancer2006;107:1520–4.

21. BeggCB,MazumdarM. Operating characteristics of a rank correlationtest for publication bias. Biometrics 1994;50:1088–101.

22. EggerM,DaveySmithG,SchneiderM,MinderC.Bias inmeta-analysisdetected by a simple, graphical test. BMJ 1997;315:629–34.

23. StataCorp. Stata/SE 11.0 for Windows. College Station, TX: StataCorporation; 2009.

24. Barrett SV, Paul J, Hay A, Vasey PA, Kaye SB, Glasspool RM. Doesbody mass index affect progression-free or overall survival in patientswith ovarian cancer? Results from SCOTROC I trial. Ann Oncol2008;19:898–902.

25. Hess LM, Barakat R, Tian C, Ozols RF, Alberts DS. Weight changeduring chemotherapy as a potential prognostic factor for stage IIIepithelial ovarian carcinoma: a Gynecologic Oncology Group study.Gynecol Oncol 2007;107:260–5.

26. Yang L, Klint A, Lambe M, Bellocco R, Riman T, Bergfeldt K, et al.Predictors of ovarian cancer survival: a population-based prospectivestudy in Sweden. Int J Cancer 2008;123:672–9.

27. Kjaerbye-Thygesen A, Frederiksen K, Hogdall EV, Glud E, ChristensenL, Hogdall CK, et al. Smoking and overweight: negative prognosticfactors in stage III epithelial ovarian cancer. Cancer Epidemiol Bio-markers Prev 2006;15:798–803.

28. Moysich KB, Baker JA, Menezes RJ, Jayaprakash V, Rodabaugh KJ,Odunsi K, et al. Usual adult bodymass index is not predictive of ovariancancer survival. Cancer Epidemiol Biomarkers Prev 2007;16:626–8.

29. Nagle CM, Purdie DM, Webb PM, Green A, Harvey PW, Bain CJ.Dietary influences on survival after ovarian cancer. Int J Cancer2003;106:264–9.

30. Schildkraut JM, Halabi S, Bastos E, Marchbanks PA, McDonald JA,Berchuck A. Prognostic factors in early-onset epithelial ovarian can-cer: a population-based study. Obstet Gynecol 2000;95:119–27.

31. Munstedt K, Wagner M, Kullmer U, Hackethal A, Franke FE. Influenceof body mass index on prognosis in gynecological malignancies.Cancer Causes Control 2008;19:909–16.

32. Skirnisdottir I, Sorbe B. Body mass index as a prognostic factor inepithelial ovarian cancer and correlation with clinico-pathologicalfactors. Acta Obstet Gynecol Scand 2010;89:101–7.

33. Suh DH, Kim HS, Chung HH, Kim JW, Park NH, Song YS, et al. Bodymass index and survival in patients with epithelial ovarian cancer. JObstet Gynaecol Res 2012;38:70–6.

34. Matthews KS, Straughn JM Jr, Kemper MK, Hoskins KE, Wang W,Rocconi RP. The effect of obesity on survival in patients with ovariancancer. Gynecol Oncol 2009;112:389–93.

35. Zhang M, Xie X, Lee AH, Binns CW, Holman CDJ. Body mass index inrelation to ovarian cancer survival. Cancer Epidemiol Biomarkers Prev2005;14:1307–10.

36. Bankhead CR, Kehoe ST, Austoker J. Symptoms associated withdiagnosis of ovarian cancer: a systematic review. BJOG 2005;112:857–65.

37. Calle EE, Kaaks R. Overweight, obesity and cancer: epidemiologicalevidence and proposedmechanisms. Nat RevCancer 2004;4:579–91.

38. ParekhN, Okada T, Lu-YaoGL.Obesity, insulin resistance, and cancerprognosis: implications for practice for providing care among cancersurvivors. J Am Diet Assoc 2009;109:1346–53.

39. Rose DP, Komninou D, Stephenson GD. Obesity, adipocytokines, andinsulin resistance in breast cancer. Obes Rev 2004;5:153–65.

40. Ronti T, Lupattelli G, Mannarino E. The endocrine function of adiposetissue: an update. Clin Endocrinol (Oxf) 2006;64:355–65.

41. Modesitt SC, van Nagell JR Jr The impact of obesity on the incidenceand treatment of gynecologic cancers: a review. Obstet Gynecol Surv2005;60:683–92.

42. PoniewierskiMS, Crawford J, Dale DC, Culakova E, Kuderer NM,WolffDA, et al. Reduced chemotherapy dose intensity in patients withovarian cancer: results from a prospective nationwide study. J ClinOncol 26: 2008 (May 20 Suppl; abstr 16508).

43. Schwartz J, Toste B, Dizon DS. Chemotherapy toxicity in gynecologiccancer patients with a body surface area (BSA)>2 m2. Gynecol Oncol2009;114:53–6.

44. Bouchardy C, Rapiti E, Blagojevic S, Vlastos AT, Vlastos G. Olderfemale cancer patients: importance, causes, and consequences ofundertreatment. J Clin Oncol 2007;25:1858–69.

45. Abrahamson PE, Gammon MD, Lund MJ, Flagg EW, Porter PL,Stevens J, et al. General and abdominal obesity and survival amongyoung women with breast cancer. Cancer Epidemiol Biomarkers Prev2006;15:1871–7.

46. DalMasoL, ZucchettoA, TalaminiR, SerrainoD,StoccoCF,VercelliM,et al. Effect of obesity and other lifestyle factors on mortality in womenwith breast cancer. Int J Cancer 2008;123:2188–94.

47. Romero-Corral A, Montori VM, Somers VK, Korinek J, Thomas RJ,Allison TG, et al. Association of bodyweight with total mortality andwith cardiovascular events in coronary artery disease: a systematicreview of cohort studies. Lancet 2006;368:666–78.

48. Smalley KJ, Knerr AN, Kendrick ZV, Colliver JA, Owen OE. Reassess-ment of body mass indices. Am J Clin Nutr 1990;52:405–8.

49. Wellens RI, Roche AF, Khamis HJ, Jackson AS, Pollock ML, SiervogelRM. Relationships between the body mass index and body compo-sition. Obes Res 1996;4:35–44.

50. Olsen CM, Green AC, Whiteman DC, Sadeghi S, Kolahdooz F, WebbPM. Obesity and the risk of epithelial ovarian cancer: a systematicreview and meta-analysis. Eur J Cancer 2007;43:690–709.

51. Schouten LJ, Rivera C, Hunter DJ, Spiegelman D, Adami HO, Arslan A,et al.Height, bodymass index, andovarian cancer: apooledanalysis of12 cohort studies. Cancer Epidemiol Biomarkers Prev 2008;17:902–12.

52. Egger M, Schneider M, Davey Smith G. Spurious precision? Meta-analysis of observational studies. BMJ 1998;316:140–4.


Protani et al.




2012;5:901-910. Published OnlineFirst May 18, 2012.Cancer Prev Res Melinda M. Protani, Christina M. Nagle and Penelope M. Webb Meta-analysisObesity and Ovarian Cancer Survival: A Systematic Review and

Updated version

10.1158/1940-6207.CAPR-12-0048doi:

Access the most recent version of this article at:

Cited articles

http://cancerpreventionresearch.aacrjournals.org/content/5/7/901.full#ref-list-1

This article cites 47 articles, 13 of which you can access for free at:

Citing articles

http://cancerpreventionresearch.aacrjournals.org/content/5/7/901.full#related-urls

This article has been cited by 3 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerpreventionresearch.aacrjournals.org/content/5/7/901To request permission to re-use all or part of this article, use this link



http://cancerpreventionresearch.aacrjournals.org/lookup/doi/10.1158/1940-6207.CAPR-12-0048

http://cancerpreventionresearch.aacrjournals.org/content/5/7/901.full#ref-list-1

http://cancerpreventionresearch.aacrjournals.org/content/5/7/901.full#related-urls

http://cancerpreventionresearch.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerpreventionresearch.aacrjournals.org/content/5/7/901


Documents

Obesity and Ovarian Cancer Survival: A Systematic …cancerpreventionresearch.aacrjournals.org/content/canprevres/5/7/...Obesity and Ovarian Cancer Survival: A Systematic Review and